Rotation Invariant Fingerprint Core-Point Detection using DWT

The performance of majority of algorithms employed for fingerprint identification is strongly affected by the accuracy of detection of core-point. Unintentional finger rotation during acquisition process inspite of mounting a finger "guide" on the sensor results in rotation of fingerprint images. In order to tackle this rotational variance a Discrete Wavelet Transform (DWT) based approach for core-point detection has been presented in this paper. The approach does not involve computation of orientation field and is directly employed on the gray scale images. We use 2D-wavelet coefficients in horizontal, vertical and diagonal direction to locate the core. Experimental results on two different databases have shown that approach